Human tuberculosis caused by Mycobacterium tuberculosis (M. tuberculosis) is a severe global health problem, responsible for approximately 3 million deaths annually, according to the WHO. The worldwide incidence of new tuberculosis (TB) cases had been falling during the 1960s and 1970s but during recent years this trend has markedly changed in part due to the advent of AIDS and the appearance of multidrug resistant strains of M. tuberculosis. 
The only vaccine presently available for clinical use is BCG, a vaccine whose efficacy remains a matter of controversy. BCG generally induces a high level of acquired resistance in animal models of TB, and in humans it is protective against disseminated forms of tuberculosis such as meningitis and miliary tuberculosis. When given to young children it is protective against tuberculosis for years but then the efficacy wanes. Comparison of various controlled trials revealed that the protective efficacy of BCG in adults varied dramatically with an efficacy range from ineffective to 80% protection. This makes the development of a new and improved vaccine against M. tuberculosis an urgent matter, which has been given a very high priority by the WHO.
Many attempts to define protective mycobacterial substances have been made, and different investigators have reported increased resistance after experimental vaccination. M. tuberculosis holds, as well as secretes, several proteins of potential relevance for the generation of a new M. tuberculosis vaccine. The search for candidate molecules has primarily focused on proteins released from dividing bacteria. Despite the characterization of a large number of such proteins only a few of these have been demonstrated to induce a protective immune response as subunit vaccines in animal models, most notably ESAT-6 and Ag85B (Brandt et al 2000). However, the demonstration of a specific long-term protective immune response with the potency of BCG or the capability of boosting in a BCG vaccinating person has not yet been achieved. At best, boost of BCG with BCG has no effect [Colditz, 1994]. Boosting of BCG has been done with Ag85A Brooks et al IAI 2001; WO0204018) in an inbred mouse strain leading to some protection, although compared to BCG alone it was not significantly better. Since BCG needs to divide and secrete proteins in order to induce a protective immune response, the lack of booster effect is primarily due to either sensitization with environmental mycobacteria or a residual immune response from the primary BCG vaccination. Both events lead to a rapid immune response against BCG and therefore quick inhibition of growth and elimination of BCG.
The course of a M. tuberculosis infection runs essentially through 3 phases. During the acute phase, the bacteria proliferate in the organs, until the immune response increases. Specifically sensitized CD4 T lymphocytes mediate control of the infection, and the most important mediator molecule seems to be interferon gamma (IFN-gamma). The bacterial loads starts to decline and a latent phase is established where the bacterial load is kept stable at a low level. In this phase M. tuberculosis goes from active multiplication to dormancy, essentially becoming non-replicating and remaining inside the granuloma. In some cases, the infection goes to the reactivation phase, where the dormant bacteria start replicating again. It has been suggested that the transition of M. tuberculosis from primary infection to latency is accompanied by changes in gene expression (Honer zu Bentrup, 2001). It is also likely that changes in the antigen-specificity of the immune response occur, as the bacteria modulates gene expression during its transition from active replication to dormancy. The full nature of the immune response that controls latent infection and the factors that lead to reactivation are largely unknown. However, there is some evidence for a shift in the dominant cell types responsible. While CD4 T cells are essential and sufficient for control of infection during the acute phase, studies suggest that CD8 T cell responses are more important in the latent phase.
In 1998 Cole et al published the complete genome sequence of M. tuberculosis and predicted the presence of approximately 4000 open reading frames (Cole et al 1998) disclosing nucleotide sequences and putative protein sequences. However importantly, this sequence information cannot be used to predict if the DNA is translated and expressed as proteins in vivo. It is known that some genes of M. tuberculosis are upregulated under conditions that mimic latency. However, these are a limited subset of the total gene expression during latent infection. Moreover, as one skilled in the art will readily appreciate, expression of a gene is not sufficient to make it a good vaccine candidate. The only way to determine if a protein is recognized by the immune system during latent infection with M. tuberculosis is to produce the given protein and test it in an appropriate assay as described herein. A number of proteins are of particular importance and have potential for being late antigens (antigens recognized during latent infection) since they are mainly expressed a relatively long time after infection where the immune system has mounted the first adaptive defense and the environment has turned more hostile for the mycobateria. In vitro hypoxic culture conditions, which mimic the conditions of low oxygen tension have previously been suggested as relevant in this regard and have been used to analyze changes in gene expression. A number of antigens have been found that are induced or markedly upregulated under these conditions e.g. the 16 kDa antigen α-crystallin (Sherman 2001), Rv2660c and Rv2659c (Betts, 2002). (our own application) Another environmental stimuli which may be of particular interest is starvation designed to reflect that nutrients are restricted in the granuloma (the location of the latent infection) and that products expressed by genes upregulated under starvation therefore may be of particular interest as antigen targets during the latent stage of infection.
Of the more than 200 hundred antigens known to be expressed during primary infection, and tested as vaccines, less than a half dozen have demonstrated significant potential. So far only one antigen has been shown to have any potential as a therapeutic vaccine (Lowrie, 1999). However this vaccine only worked if given as a DNA vaccine and has proved controversial, with other groups claiming that vaccination using this protocol induces either non-specific protection or even worsens disease (Turner, 2000). In contrast, the fusion polypeptides described in the invention may be incorporated in a vaccine that use well-recognized vaccination technology, as demonstrated in provided examples.
Further, since TB vaccines do not result in sterilizing immunity but rather control the infection at a subclinical level (thereby resulting in the subsequent establishment of latent infection), a multiphase vaccine which combines components with prophylactic and therapeutic activity is described in this invention. After conventional prophylactic vaccination, the evasion of the primary immune response and the subsequent development of latent disease is probably at least in part due to the change in the antigenic profile of the invading bacteria. Thus, vaccinating with antigens associated with latent TB should prevent or reduce the establishment of latent infection and therefore, a vaccine incorporating antigens expressed by the bacteria both in the first logarithmic growth phase and during latent disease should improve long-term immunity when used as a prophylactic vaccine. Such a multiphase vaccine will obviously also be efficient as a therapeutic vaccine thereby addressing the problem that the majority of the population in the third world who would receive a future TB vaccine would be already latently infected.